Xu et al. and Disease (2021) 12:406 https://doi.org/10.1038/s41419-021-03668-x Cell Death & Disease

ARTICLE Open Access

Malignant pleural mesothelioma co-opts BCL-XL and to escape Duo Xu1,2, Shun-Qing Liang1,2, Zhang Yang1,2, Haitang Yang 1,2,RémyBruggmann 3, Simone Oberhaensli3, Sabina Berezowska 4,ThomasM.Marti1,2,SeanR.R.Hall1,2, Patrick Dorn1,2,GregorJ.Kocher1,2,RalphA.Schmid1,2 and Ren-Wang Peng 1,2

Abstract Escape from is a hallmark of cancer. In this study, we investigated the anti-apoptotic mechanisms and explored the therapeutic potential of BCL-2 domain-3 (BH3) mimetics in malignant pleural mesothelioma (MPM), a lethal thoracic malignancy with an extreme dearth of treatment options. By implementing integrated analysis of functional genomic data of MPM cells and quantitative proteomics of patients’ tumors, we identified BCL-XL as an anti-apoptotic driver that is overexpressed and confers an oncogenic dependency in MPM. MPM cells harboring genetic alterations that inactivate the NF2/LATS1/2 signaling are associated with increased sensitivity to A-1155463, a BCL-XL-selective BH3 mimetic. Importantly, BCL-XL inhibition elicits protective autophagy, and concomitant blockade of BCL-XL and autophagic machinery with A-1155463 and hydroxychloroquine (HCQ), the US Food and Drug Administration (FDA)-approved autophagy inhibitor, synergistically enhances anti-MPM effects in vitro and in vivo. Together, our work delineates the molecular basis underlying resistance to apoptosis and uncovers an evasive mechanism that limits response to BH3 mimetics in MPM, suggesting a novel strategy to target this aggressive disease. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,;

Introduction in MPM have revealed frequent oncogenic events Malignant pleural mesothelioma (MPM) is a highly enabled by genetic alterations that inactivate tumor aggressive malignancy that is etiologically associated suppressor , most often BRCA1 associated - with asbestos exposure1,2. Despite the restriction of 1(BAP1), neurofibromatosis type 2 (NF2), large tumor asbestos use in most countries, the incidence of MPM is suppressor kinase 2 (LATS2), and cyclin-dependent still rising due in part to the long latency (around 40 kinase inhibitor 2 A/2B (CDKN2A/2B), which, however, years) of the interval from carcinogen exposure to tumor have proven difficult to be therapeutically exploited2,5,6. onset3. There are no typical clinical symptoms of Further exacerbating the dilemma, platinum-based che- mesothelioma in the early phase, and the majority of motherapy, the current standard of care for inoperable patients (80%) are diagnosed at advanced stages asso- late-stage MPM, only marginally improves patient sur- ciated with extremely poor prognosis4. Previous studies vival7. Hence, there is a pressing need to identify new druggable targets in MPM and develop effective ther- apeutic strategies for the daunting disease. Correspondence: Ralph A. Schmid ([email protected]) or Ren- The NF2 tumor suppressor encodes Merlin Wang Peng ([email protected]) (Moesin-ezrin-radixin-like protein), which mediates 1Division of General Thoracic Surgery, Inselspital, Bern University Hospital, Bern, Switzerland tumor suppression and contact-dependent inhibition by 2Department for BioMedical Research (DBMR), University of Bern, Bern, repressing Hippo, mTORC1, RAS, EGFR, and FAK-Src Switzerland signaling pathways8. The Hippo signaling, an evolu- Full list of author information is available at the end of the article These authors contributed equally: Duo Xu, Shun-Qing Liang, Zhang Yang tionally conserved pathway that regulates organ size and Edited by N. Barlev

© The Author(s) 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a linktotheCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 2 of 14

tissue homeostasis by restricting cell growth and pro- regularly tested free of mycoplasma (Microsynth). A- moting apoptosis, is one of the best characterized Merlin/ 1155463 (Cat. #CS-5398), hydroxychloroquine (Cat. NF2-regulated pathways9. Besides the mutation in NF2, #CS-8017), Venetoclax (Cat. #S8048), and ABT-263 other components of the Hippo pathway, e.g., large tumor (Cat. #S1001) were obtained from ChemScene and suppressor kinase 1/2 (LATS1/2), are also frequently SelleckChem, respectively. Bafilomycin A1 was provided inactivated in MPM patients6. Dysregulation of the Hippo by K. Krempaska (Department for Biomedical Research, pathway constitutively activates Yes-associated protein University of Bern, Switzerland). (YAP), a transcription regulator that promotes the tran- scription of genes involved in cell proliferation and anti- Cell viability and clonogenic survival assay apoptosis by interaction with TEA/ATTS domain (TEAD) Cell viability was measured by acid phosphatase (APH) transcription factors10. assay as described18,20. Each data point was generated in Resistance to apoptosis, a critical barrier of tumor triplicate from three independent experiments (n = 3). development, is one of the most prominent hallmarks of IC50 values were determined based on the best-fit curve cancer11. Overexpression of pro-survival B-cell lymphoma generated in GraphPad Prism [log (inhibitor) vs. nor- 2 (BCL-2) family members (BCL-2, BCL-XL, MCL-1, malized response]. Combination Index (CI) was calcu- BCL-W,BCL-B, and BFL-1) is a key apoptosis evasion lated by ComboSyn software21. CI < 1.0, synergism; CI = mechanism that promotes tumor growth by keeping pro- 1.0, additive effects, CI > 1.0, antagonism. Clonogenic apoptosis effectors (BAX/BAK) in check11. By contrast, the assay was done as described18,20. Briefly, cells seeded in BCL-2 homology domain-3 (BH3)-only (BAD, 6-well plates (1000–2000 cells/well) were treated for 96 h BIM, BID, NOXA, PUMA, BIK, BMF, and HRK) induce and cultured in the absence of drugs for 10–12 days apoptosis by neutralizing the pro-survival BCL-2 pro- depending on growth rate. The resulting colonies teins12. As such, targeting anti-apoptotic regulators with were stained with crystal violet (0.5% dissolved in 25% BH3 mimetics represents an attractive strategy for cancer methanol). 13 therapy . Several BH3 mimetics, e.g., the BCL-2/BCL-XL/ BCL-W inhibitor ABT-263 (navitoclax), BCL-2–selective Immunoblotting and immunohistochemistry inhibitor venetoclax (ABT-199), and BCL-XL–selective Cell lysates were prepared and immunoblot analysis was inhibitor A-1155463, have showed promising clinical performed as described18,22. In brief, protein lysates were activity14. In particular, venetoclax has been approved by resolved by SDS-PAGE (Cat. #4561033; Bio-Rad Labora- the US Food and Drug Administration (FDA) for the tories) and transferred onto nitrocellulose membranes treatment of chronic lymphocytic leukemia (CLL) with a (Cat. #170-4158; Bio-Rad). After incubation with blocking 17p-deletion or TP53 mutation15. We and others have buffer (Cat. #927-4000; Li-COR Biosciences) for 1 h at reported that MPM cells can acquire anti-apoptotic room temperature, membranes were incubated with pri- adaptation as a protective mechanism to evade onco- mary antibodies (BCL-XL: 1:1000, #2764; Cleaved Cas- – genic stress and anticancer therapy16 18. In this study, we pase-7: 1:1000, #9491; LC3B: 1:500, #12741; Beclin-1: systematically analyzed the cell survival dependency on 1:1000, #3495; p62: 1:500, #5114; ATG5:1:1000, #12994; anti-apoptotic BCL-2 proteins and explored the potential Cell Signaling Technology) overnight at 4 °C. IRDye of specific BH3 mimetics as anti-MPM therapy. 680LT-conjugated goat anti-mouse IgG (Cat. #926- 68020) and IRDye 800CW-conjugated goat anti-rabbit Materials and methods IgG (Cat. #926-32211) from Li-COR Biosciences were Cell culture and reagents used at 1:10000 dilutions. Signals of membrane-bound Human normal mesothelial cells (LP-9) was a gift from secondary antibodies were visualized by the Odyssey Robert Kratzke (Masonic Cancer Center, University of Infrared Imaging System (Li-COR Biosciences), followed Minnesota, USA)19 and cultured in Medium 199 (Cat. by quantification using Image J23. #M7528; Sigma-Aldrich) supplemented with 15% fetal Immunohistochemical study were performed as bovine serum (Cat. #10270-106; Life Technologies), 1% described24. In brief, surgically removed xenograft tumors penicillin/streptomycin solution (Cat. #P0781, Sigma- (two tumors/group) were formalin-fixed, paraffin- Aldrich), 10 ng/ml of epidermal growth factor (Cat. embedded (FFPE), and stained with hematoxylin and #E5036, Sigma-Aldrich), and 0.4 μg/ml hydrocortisone eosin (H&E) using standard protocols. FFPE tissue blocks (Cat. #07904, STEMCELL TECHNOLOGIES) at 37 °C were sectioned at 4 μm, deparaffinized, rehydrated, and with 95% air/5% CO2. Human normal lung fibroblasts subsequently stained with appropriate antibodies (LC3B: (hFb16Lu;CCD-16Lu) and human MPM cell lines 1:4000, #3868, Cell Signaling Technology; p62: 1:8000, # (MESO-1, MESO-4, JL-1, H2452, MSTO-211H, and WH0008878M1, Sigma; Cleaved Caspase-3: 1:200; # H2052) were previously described18,20. All cells were 9664, Cell Signaling Technology) using the automated authenticated by short tandem repeat (STR) profiling and system BOND RX (Leica Biosystems)24. Visualization was

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 3 of 14

Fig. 1 (See legend on next page.) performed using the Bond Polymer Refine Detection kit Apoptosis assay (Leica Biosystems) as instructed by the manufacturer. MPM cells were treated as specified in the figure legends. ® Images were acquired using PANNORAMIC whole slide After treatment, cells in the supernatant and adherent to scanners and processed using Case Viewer (3DHISTECH plates were collected, washed with PBS, and pooled before Ltd.). The staining intensities of the whole slide (two suspended in 400 μl binding buffer and stained with the tumors/group) were quantified by QuPath software25. Annexin V Apoptosis Detection Kit-FITC (Cat. #88-8005;

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 4 of 14

(see figure on previous page)

Fig. 1 BCL-XL is overexpressed and confers a survival dependency in MPM. A Genetic alterations and mRNA/protein expression of pro-survival Bcl-2 family genes in the TCGA cohort of MPM patients (n = 87). Data were downloaded from the cBio Cancer Genomics Portal (CBioPortal). B, C The dependency profile of pro-survival BCL-2 family genes (BCL2L1, BCL2L10, BCL2, BCL2A1, BCL2L2, and MCL1) in MPM cell lines based on genome-wide CRISPR (NCIH28, NCIH2452, MPP89, NCIH2052, ISTMES2, MERO14, and MSTO-211H) or RNAi interference (MPP89, ACCMESO-1, NCI-H28, NCIH2452, NCI- H2052, and JL-1) screens. Data were downloaded from The Cancer Dependency Map Project (DepMap) datasets: CRISPR (Avana) Public 19Q3 and Combined RNAi (Broad, Novartis, Marcotte). D Volcano plot of transcriptomic comparison between patients’ MPM samples (n = 40) versus normal pleural tissues (n = 5) (GSE2549). Genes significantly downregulated (adjusted p value <0.05) are shown in green and genes significantly upregulated

in red. The differentially expressed pro-survival Bcl-2 family genes are highlighted. E Immunoblots of BCL-XL in human normal fibroblasts (hFb16Lu), normal mesothelial cells (LP-9), and MPM cells. Proteins lysates prepared from the cells were subjected to serial dilutions (100, 75, and 50%) for

immunoblot analysis (left). Densitometric analysis of the immunoblot (right) showed fold change of BCL-XL signal normalized against Actin and the total protein, with the value of hFb16Lu cells set to 1. F BCL-XL protein level in TCGA pan-cancer cohort (n = 32). The expression profile was obtained from The Cancer Proteome Atlas (TCPA). READ rectum adenocarcinoma, CHOL cholangiocarcinoma, MESO mesothelioma, COAD colon adenocarcinoma, LUAD lung adenocarcinoma, SKCM skin cutaneous melanoma, LUSC lung squamous cell carcinoma, THCA thyroid carcinoma, UCEC uterine corpus endometrial carcinoma, PAAD pancreatic adenocarcinoma, GBM multiforme, KICH kidney chromophobe, BLCA bladder urothelial carcinoma, HNSC head and neck squamous cell carcinoma, PRAD prostate adenocarcinoma, OV ovarian serous cystadenocarcinoma, LGG brain lower grade glioma, THYM thymoma, STAD stomach adenocarcinoma, DLBC lymphoid neoplasm diffuse large B-cell lymphoma, TGCT testicular germ cell tumors, BRCA breast invasive carcinoma, KICH kidney chromophobe, KIRP kidney renal papillary cell carcinoma, UVM uveal melanoma, UCS uterine carcinosarcoma, CESC cervical squamous cell carcinoma and endocervical adenocarcinoma, ESCA esophageal carcinoma, PCPG pheochromocytoma and paraganglioma, SARC sarcoma, LIHC liver hepatocellular carcinoma, ACC adrenocortical carcinoma.

Thermo Fisher Scientific) according to the manufacturer’s but not blinded manner. Suspensions of MESO-1 cells instructions. Flow cytometry analysis was performed on a mixed 1:1 with Matrigel (Cat. #356231; Corning) were BD Biosciences LSRII flow cytometer. subcutaneously inoculated in the flanks (1 × 106 cells /injection). One month after injection, mice were randomly Autophagic flux assay assigned to treatment groups (n = 5): (1) control; (2) A- The mCherry-eGFP-LC3B lentivirus was kindly pro- 1155463 (5 mg/kg, i.p., once daily); (3) HCQ (50 mg/kg, i.p., vided by Mario P. Tschan (Institute of Pathology, Uni- once daily); (4) combination of A-1155463 and HCQ in the versity of Bern, Switzerland). Briefly, lentivirus was abovementioned doses(Tumor size was measured by digital transduced into cell lines followed by selection with caliper every two to three days. Tumor volume was cal- puromycin (1 μg/ml) for 3 days and various treatments. culated as follows: (length x width x width)/2. Mice were Cells were then trypsinized and resuspended for flow sacrificed at the end of 23-day treatment. cytometry analysis of GFP and mCherry fluorescence using a BD Biosciences LSRII flow cytometer. Data were Public databases (TCGA, TCPA, CBioPortal, DepMap, GEO, analyzed by FlowJo software and gates for populations GSDC, and COSMIC) with low/intermediate/high mCherry/GFP ratio were set Interrogation of publicly available dataset was per- according to previous studies26,27. formed as we have described28. Specifically, transcriptome profiling and reverse-phase protein array data of meso- Small interfering RNA (siRNA) knockdown thelioma patients were obtained from the Cancer Genome Knockdown of BCL2L1 and ATG5 was achieved by Atlas (TCGA), the Cancer Proteome Atlas (TCPA), and specificduplexsiRNAs(BCL2L1 siRNA, 15 nM; ATG5 the cBio Cancer Genomics Portal (CBioPortal)29,30. The siRNA, 30 nM) purchased from Origene Technologies catalog of gene essentiality across MPM cell lines is (Cat. #SR319459 and SR322789). Transfection of siR- obtained from the Cancer Dependency Map Project NAs was performed with Lipofectamine 2000 (Cat. (DepMap)31. Transcriptomic data of MPM samples #116628027, Invitrogen) according to the manu- (GSE2549) was downloaded from the facturer’s instructions. Omnibus(GEO)32. Genomics of Drug Sensitivity in Can- cer (GDSC) and Catalog of Somatic Mutations in Cancer Animal experiments (COSMIC) was used to extract transcriptomic data and Mouse experiments were conducted in accordance with mutation status of MPM cell lines33,34. Institutional Animal Care and Ethical Committee- approved animal guidelines and protocols. Experiments Statistical analysis were performed in 8-week-old male NSG (NOD-scid Statistical analyses were performed using GraphPad null IL2Rγ ) mice, with sample size not predetermined by Prism 8 (GraphPad Software, Inc.). All samples that met statistical method but rather based on preliminary experi- proper experimental conditions were included in the ments. Group allocation was performed in a randomized analysis. Data represent biological replicates (n) and are

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 5 of 14

Fig. 2 (See legend on next page.)

depicted as mean values ± s.d. or mean values ± SEM as t-test, one-way or two-way ANOVA as indicated in the indicated in the figure legends. Comparison of mean figure legends, P < 0.05 were considered statistically values was conducted with unpaired, two-tailed Student’s significant.

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 6 of 14

(see figure on previous page)

Fig. 2 Genetic and pharmacological inhibition of BCL-XL preferentially impairs MPM cell proliferation. A-C Immunoblots (A) and micrographic images (B) of LP-9 and MPM cells with siRNA-based BCL2L1 knockdown. Viable cells were counted 48 h post-transfection by trypan blue dye exclusion (C). Data were presented as mean ± s.d. (n = 3). ****p < 0.0001 by two-way ANOVA with Tukey’s multiple comparisons test. Scale bar: 50 μm. D Flow cytometry-based apoptotic assay of LP-9 and MESO-1 cells 48 h post-transfection with siRNAs. Shown on the right is fold change in apoptotic cells (Annexin V-positive) induced by BCL2L1 knockdown (siBCL2L1) compared with control siRNA (siCtrl), with the value of the control treatment set as 1. Data were presented as mean ± s.d. (n = 3), with a representative plot shown in the left. **p < 0.01 by two-way ANOVA with Tukey’s multiple comparisons test. E Clonogenic assay of LP-9 and MPM cells treated with A-1155463 for 96 h and subsequently cultured in the absence of the drug for additional 12 days. Quantification is shown underneath, with data presented as mean ± s.d. (n = 2). F Immunoblots of LP-9 and MESO-1 cells treated with A-1155463 for 24 h. G Flow cytometry-based apoptotic assay of LP-9 and MESO-1 cells after treatment with A-1155463

for 48 h. Shown on the right is fold change in apoptotic cells (Annexin V-positive) induced by treatment with A-1155463 treatment (BCL-XLi) compared with vehicle control (0 nM), with the value of the control treatment set as 1. Data were presented as mean ± s.d. (n = 3), with a representative plot shown in the left. ***p < 0.001 by two-way ANOVA with Tukey’s multiple comparisons test.

14 Results BCL-XL inhibitor , preferentially impaired MPM cells, BCL-XL is deregulated and confers a survival dependency resulting in significantly greater growth inhibition in in MPM MPM cells than in LP-9 cells (Fig. 2E). Importantly, In the attempt to explore the potential of BH3 mimetics A-1155463 induced cleavage of caspase-7 (Fig. 2F) and a as anti-MPM therapy, we investigated genetic status, dose-dependent increase of apoptotic cells in MESO-1, as transcriptional expression, and dependency profile of the manifested by flow cytometry-based apoptotic analysis, pro-survival BCL-2 gene family (BCL2L1, BCL2L10, which showed that treatment with 62.5, 125, and 250 nm BCL2, BCL2A1, BCL2L2, and MCL1) in MPM by inter- A-1155463 resulted in 4-, 4.4-, and 5.4-fold increases in rogating TCGA dataset and functional genomics that apoptotic cells (Annexin V-positive) compared with determines genetic dependencies in cancers (DepMap; vehicle treatment (Fig. 2G). In contrast, A-1155463 https://depmap.org/portal/). The integrated molecular treatment barely increased apoptosis compared with characterization revealed that, of the anti-apoptotic genes, vehicle control in LP-9 cells (Fig. 2F, G). Importantly, the BCL2L1 (encoding BCL-XL) is altered in a subset of MPM selective BCL-2 inhibitor Venetoclax failed to distinguish patients (n = 87), by means of gene amplification and malignant from normal mesothelial cells, leading to mRNA overexpression (Fig. 1A) and MPM cells show the almost equal effects on MPM and LP-9 cells (Supple- greatest dependency on BCL2L1 for survival (Fig. 1B, C). mentary Fig. 2A). Taken together, these results reveal that Consistently, BCL2L1 expression is significantly upregu- BCL-XL is highly deregulated and confers an oncogenic lated in patients’ MPM compared with that in normal dependency in MPM. pleural tissues (Fig. 1D). Our immunoblot analysis revealed upregulated expression of BCL-XL in human NF2/LAST1/2 mutations are associated with increased MPM cell lines compared with normal lung fibroblasts sensitivity to BCL-XL inhibition in MPM (hFb16Lu) and mesothelial LP-9 cells (Supplementary Next, we sought to identify potential biomarkers asso- Fig. 1). Notably, a remarkably greater increase in BCL-XL ciated with MPM response to BCL-XL inhibition in MPM. was observed in MPM cells compared with normal con- As expected, the BCL-XL protein level was positively trols when BCL-XL signal was normalized against correlated with the sensitivity to A-1155463 [negatively Actin (loading control) and the total protein (Fig. 1E). with the IC50 (50% inhibitory concentration)] in MPM Further supporting these observations, examination cells (Fig. 3A, B). of TCPA dataset, which provides quantitative proteomics It has been shown that cancer cells express high levels of of patient-derived pan-cancers (n = 32), revealed that pro-apoptotic proteins that, however, are constrained via MPM had the third highest level of BCL-XL (Fig. 1F). heterodimerization by anti-apoptotic effectors during 11,35 These results indicate that anti-apoptotic BCL-XL is tumorigeneis . As a result, cancer cells can be con- deregulated in MPM at genetic, transcriptional, and sidered to be ready to undergo apoptosis or “primed for translational levels. apoptosis”, highlighting the potential of BH3 mimetics in Next, we addressed whether BCL2L1 represents a the clinic36. To explore a possible link of BCL2L1 with genetic vulnerability in MPM. BCL2L1 knockdown by pro-apoptotic proteins in MPM, we examined a cohort of siRNAs caused significantly greater proliferative inhibi- MPM patients (n = 87) in TCGA, which revealed that tion and apoptotic cell death in MPM cells (MESO-1, BCL2L1 mRNA level was positively correlated with that of MESO-4, JL-1, H2452, MSTO-211H, H28, and H2052) several pro-apoptotic genes, e.g., BAX, BBC3, BIK, and than in LP-9 cells (Fig. 2A–D). Consistent with the BAK1 (Fig. 3C). Consistent with this observation, MPM genetic results, A-1155463, a potent and highly selective cells that overexpress BCL-XL also have higher protein

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 7 of 14

Fig. 3 (See legend on next page.) levels of BAX (Supplementary Fig. 2B), suggesting that expression of the pro-apoptotic genes (BAK1, BAX, BBC3, BCL2L1-positive MPM tumors are “primed” to apoptosis BID, BIK, BOK, and HRK)37, we curated the “primed for induction. By using a previously defined “primed for apoptosis” score of H2452, MSTO-211H, H28, and H2052 apoptosis” gene signature, determined by transcriptional cells, chosen for the availability of their transcriptomic

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 8 of 14

(see figure on previous page)

Fig. 3 NF2/LAST1/2 mutations are associated with increased sensitivity to BCL-XL inhibition in MPM cells. A Cell viability assay of LP-9 and a panel of MPM cell lines treated with A-1155463 for 96 h. Data were presented as mean ± s.d. (n = 3). B Correlation analysis of BCL-XL protein levels and the 50% inhibitory concentration (IC50) values of A-1155463 in MPM cell lines. C Volcano plot of Spearmans’ rank correlation coefficient between gene expression of BCL2L1 and that of pro-apoptotic genes in patients’ MPM samples (n = 87). Data were downloaded from TCGA and subsequently analyzed by R software (Cor.test function). D Correlation analysis of “primed” for apoptosis signature (based on the sum of gene expression of BAK1,

BAX, BBC3, BID, BIK, BOK, and HRK) and the IC50 of A-1155463 in MPM cell lines. Transcriptomic data of the indicated MPM cell lines were extracted from Catalog of Somatic Mutations in Cancer (COSMIC). E Genetic status of NF2/LATS1/2 in TCGA cohort of MPM patients (n = 87). Data were downloaded from cBioPortal. F Genetic alterations in NF2/LATS1/2 are associated with decreased protein level of YAP1_pS127. MPM patients (n = 61) in a TCGA cohort were stratified according to NF2/LATS1/2 genetic status (altered or unaltered), with p value calculated by unpaired, two-tailed Student’s t-test. G Kaplan–Meier curves showing overall survival (OS) of a TCGA cohort of MPM patients (n = 61) stratified by protein level of YAP1_pS127 (high versus low). The p value is calculated by the log-rank test using R (version 3.4.3). H NF2/LATS1/2-mutant MPM cells showed . increased sensitivity to A-1155463 (IC50) MPM cells were grouped according to genetic status of NF2/LATS1/2, with p value calculated by unpaired, two-tailed Student’s t-test. WT, no alterations in NF2/LATS1/2. I Genetic alterations in NF2/LATS1/2 are associated with increased BCL2L1 mRNA levels and with a higher gene signature “primed” for apoptosis. MPM patients of a TCGA cohort (n = 87) were grouped according to NF2/LATS1/2 genetic status in the tumors (altered or unaltered). p value was calculated by unpaired, two-tailed Student’s t-test. data in the Catalog of Somatic Mutations in Cancer suggesting a possible role for autophagy to protect MPM (COSMIC). Our analysis revealed that MPM cells with cells from the stress elicited by BCL-XL inhibition. Sup- higher scores of the apoptosis gene signature were more porting our hypothesis, treatment with A-1155263 acutely sensitive (lower IC50 values) to A-1155463 (Fig. 3D), increased autophagic activity in MESO-1 cells, marked by supporting the notion that expression of the pro- decreased p62, upregulated Beclin-1, and microtubule- apoptotic genes predicts BCL-XL inhibitor sensitivity. associated proteins 1A/1B light chain 3B (LC3B)-II Finally, we explored if recurrent genetic alterations in (Fig. 4B), whereby the conversion of cytosolic LC3B-I to MPM are associated with sensitivity to BCL-XL inhibition. autophagosome-localized LC3B-II is proportional with NF2/LATS1/2 loss of function (deletion, truncation, and initiation of autophagy and therefore serves as a reliable mutation) is frequent (38%) in MPM (Fig. 3E), which marker of autophagosomes38. Moreover, genetic (siRNAs) downregulates YAP phosphorylation (YAP1_pS127) and and pharmacological (A-1155463) inhibition of BCL-XL increases the activity of YAP oncoprotein (Fig. 3F). Con- markedly increased the LC3B-II lapidated form, in parti- sistent with this observation, MPM patients with low cular the ratio of LC3B-II/Actin signal, compared with levels of YAP1_pS127 (increased YAP activity) were vehicle controls at the basal level and in the presence of associated with dismal prognosis (Fig. 3G) and MPM cells Bafilomycin A1, an inhibitor of autophagosome-lysosome harboring NF2/LATS1/2 mutations/deletions exhibited fusion26,38, in a panel of MPM cells (Fig. 4C–F). In sharp increased sensitivity to A-1155463 (Fig. 3H). Supporting contrast, the same effects of BCL-XL inhibition were not this finding, examining the TCGA cohort of MPM observed in LP-9 cells, as A-1155463 alone showed no patients (n = 87) showed that NF2/LATS1/2-altered effect on the LC3B-II/Actin ratio (compared with vehicle tumors were characterized by an increased BCL2L1 control), as did concomitant treatment with A-1155463 expression and “primed for apoptosis” score (Fig. 3I). and Bafilomycin (compared with Bafilomycin alone) Thus, BCL-XL protein expression, “primed for apopto- (Supplementary Fig. 2C). Importantly, using the mCherry- sis” gene signature and genetic alterations in NF2/LATS1/ eGFP-LC3B fluorescence reporter26,27, we showed that 2 may serve as biomarkers to stratify MPM subsets that genetic and pharmacological inhibition of BCL-XL sig- likely benefit from BCL-XL targeted therapy. nificantly increased the mCherry:GFP fluorescence ratio, a well-recognized measure of autophagic flux, in MPM cells

BCL-XL inhibition elicits protective autophagy in MPM cells stably expressing mCherry-eGFP-LC3B (Fig. 4G, H), fur- Our observations that MPM cells show heterogeneous ther strengthening the notion that BCL-XL inhibition responses to BCL-XL inhibition (Figs. 2, 3) suggest the increases autophagy. existence of resistance mechanisms. This prompted us to SQSTM1/p62 deregulation (e.g., mRNA upregulation or explore the approaches to improve the efficacy of BCL-XL- protein overexpression) occurred in a subset (9%) of targeted therapy. Interrogation of TCPA dataset revealed MPM patients (Fig. 4I). Although implications of SQSTM1 (sequestosome 1; also termed p62) as the top SQSTM1/p62 changes can be context-dependent and candidate that is significantly negatively correlated with deserve cautious interpretations, the steady state level of 38 BCL-XL (Fig. 4A). p62 is a key component of autophagic p62 do reflect the autophagic status and it is widely machinery functioning as a cargo adapter by physical accepted that impaired autophagy contributes to initiation interaction with and subsequent delivery of autophagic and early development of cancer38,39. Indeed, previous substrates to autophagosomes for degradation38, studies associated decreased p62 or high autophagy status

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 9 of 14

Fig. 4 (See legend on next page.)

with better clinical outcomes in MPM and other prognostic value of BCL-XL (Fig. 4J, Supplementary Fig. tumors40,41. Supporting this notion, high p62 levels pre- S3A). Moreover, deregulation of SQSTM1/p62 and dicted poorer prognosis in MPM patients, opposite to the BCL2L1 appeared mutually exclusive (Fig. 4I), reiterating

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 10 of 14

(see figure on previous page)

Fig. 4 Genetic and pharmacological inhibition of BCL-XL increases autophagic flux in MPM cells. A, Volcano plot of Spearmans’ rank correlation coefficient between BCL-XL and other tested proteins (n = 217). Proteomic data were downloaded from TCPA, and proteins significantly (adjusted p value (or q value) <0.05) correlated with BCL-XL are highlighted. B Immunoblots of MESO-1 treated with A-1155463 for 24 h. C, D MPM cells transiently transfected with BCL2L1-siRNAs (siBCL2L1) were treated (48 h post-transfection) with Bafilomycin (200 nM) or vehicle (DMSO) for 2 h before subjected to immunoblots to assess autophagic flux. Protein quantification is shown in (D). Note the increase in LC3B-II/Actin ratio by BCL2L1 knockdown in the absence and presence of Bafilomycin. E, F MPM cells pre-treated with A-1155463 (1 μM) for 24 h were exposed to Bafilomycin (200 nM) or vehicle (DMSO) for another 2 h before immunoblot analysis. Protein quantification is shown in (F). Note the increase in LC3B-II/Actin ratio by A-1155463 treatment in the absence and presence of Bafilomycin. G, H MPM cells stably expressing mCherry-eGFP-LC3B were transfected with BCL2L1-siRNA (siBCL2L1) or control siRNA for 48 h (G), or treated with A-1155463 (1 μM) or vehicle for 24 h (H) before analyzed by flow cytometry. Data were presented as mean ± s.d. (n = 3). *p < 0.05, **p < 0.01, and ****p < 0.0001 by two-way ANOVA with Sidak’s multiple comparisons test of control high versus experimental high (red). I Genetic status and mRNA/protein expression of BCL2L1 and SQSTM1 in a TCGA cohort of MPM patients (n = 87). Data were downloaded from cBioPortal. J Kaplan–Meier analysis of MPM based on p62 protein level. TCGA cohort of MPM patients (n = 61) with high- (in red) or low-p62 (in black) were stratified by optimal cutoff value of the p62 across all patients using the surv_cutpoint function in the R “maxstat” package. The p value is calculated by the log-rank test using R (version 3.4.3).

the reciprocal nature of p62-and BCL-XL-regulated pro- alone delayed tumor growth, the addition of HCQ pro- cesses in MPM development. Together, these results foundly enhanced antitumor efficacy without obvious side reveal that BCL-XL inhibition elicits autophagy, which effects, i.e., body weight loss (Fig. 6A–D). Immunohisto- may act as a compensatory mechanism that counteracts chemical analysis demonstrated that tumors treated with BCL-XL targeted therapy. A-1155463 alone showed enhanced punctate staining of LC3B (highlighted in insets) and reduced p62 compared

Concomitant blockage of BCL-XL and autophagy with the vehicle group (Fig. 6E), consistent with an increase synergistically enhances anti-MPM effects in the autophagic flux upon BCL-XL inhibition as we To test this hypothesis that autophagy protects MPM showed in vitro. Importantly, the combination treatment cells from the cytotoxicity of BCL-XL inhibition, MPM blunted A-1155463-elicited autophagy, accompanied by cells were concomitantly treated with A-1155463 and the increase in tumor cell apoptosis as indicated by the increase autophagy inhibitor hydroxychloroquine (HCQ) across a in cleaved caspase-3 in the combination group compared broad range of concentrations. While single agents sup- with single treatment (Fig. 6E). Immunoblots indicated that pressed cell proliferation in a dose-dependent manner, A- drug combination of A-1155463 and HCQ increased p62 1155463 plus HCQ produced a strong synergy, leading to and cleaved caspase-7 compared with single agents alone significantly enhanced antiproliferative effects and apop- (Fig. 6F), further supporting the notion that combined totic cell death in a panel of MPM cells, including NF2/ A-1155463 and HCQ suppresses autophagy and induces LATS1/2-mutant and wild-type (Fig. 5A–E) but not in apoptosis. Overall, these in vivo data validate a novel ther- human normal lung fibroblasts (Supplementary Fig. S3B). apeutic strategy by combined inhibition of BCL-XL and The synergy also applied when HCQ was combined with autophagy to target MPM. ABT-263, a pan-inhibitor against BCL-2, BCL-XL, and BCL-W (Supplementary Fig. S3C–E). Discussion To further explore the role of autophagy in MPM In this study, we showed that MPM capitalizes on BCL- response to BCL-XL inhibition, we knocked down XL for anti-apoptosis, and that NF2/LATS1/2-alterations autophagy related 5 (ATG5), which encodes a key effector and pro-apoptotic gene expression are associated with protein (ATG5) involved in the initiation of pre- sensitivity to BCL-XL inhibition. We further revealed that 39 autophagosome formation . Genetic depletion of ATG5 BCL-XL blockage elicited protective autophagy, such that in MPM cells significantly enhanced the antiproliferative combined treatment with BCL-XL-selective BH3 mimetic effects of BCL-XL inhibition (Fig. 5F, G), which is in line and clinically approved autophagy inhibitor yields strong with the results of pharmacological studies (Fig. 5A–D). and synergistic anti-MPM effects in vitro and in vivo. Our Thus, BCL-XL inhibition elicits protective autophagy and data suggest the therapeutic potential of targeting BCL-XL combined blockage of BCL-XL and autophagy represents alone for MPM subsets, and of co-targeting autophagy for a promising strategy to treat MPM. unselected MPM. Apoptosis is regulated by pro- and anti-apoptotic BCL-2 11,13 BCL-XL inhibition combined with hydroxychloroquine proteins, which is invariably deregulated in cancer .In potently suppresses MPM growth in vivo response to oncogenic or stress signals, malignant cells To extend the in vitro observations to in vivo, we eval- overexpress anti-apoptotic proteins to dampen apoptosis uated efficacy of the combination treatment with A-1155463 by sequestering pro-apoptotic activators35. In this sce- and HCQ in MESO-1 xenografts. Whereas A-1155463 nario, cancer cells are proposed to be “primed” for

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 11 of 14

Fig. 5 Dual inhibition of BCL-XL and autophagy synergistically enhances antiproliferative effects and increases apoptosis. A, B Dose- response curves (A) of MPM cells treated with A-1155463 and HCQ, alone or in combination, for 96 h. Combination index (CI) values (B) were determined by the CompuSyn software. Data were shown as mean ± s.d. (n = 3). CI < 1.0, synergism; CI = 1.0, additive effects, CI > 1.0, antagonism. C Clonogenic assay of MPM cells treated for 96 h with A-1155463 and HCQ, alone or in combination. Cells were cultured without the drug for additional 12 days before crystal violet staining. Representative images of three independent experiments (n = 3) are shown. D Flow cytometry- based apoptosis assay of MPM cells treated with A-1155463 and HCQ, alone or in combinations for 48 h. Data were presented as mean ± s.d. (n = 3), with a representative plot shown in the left. *p < 0.05, ** p < 0.01, ***p < 0.005, and **** p < 0.001 by one-way ANOVA with Sidak’s multiple comparisons test. E Immunoblots of MESO-1 cells treated with A-1155463 and HCQ, alone and in combination, for 24 h. F, G MPM cells transfected with ATG5- or control siRNAs (siATG5, siCtrl) were analyzed (48 h post-transfection) by immunoblots (F) or treated with A-1155463 for another 24 h, followed by quantification of viable cells (G). Data were shown as mean ± s.d. (n = 3). *p < 0.05 and ** p < 0.01 by unpaired t-test.

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 12 of 14

Fig. 6 Hydroxychloroquine (HCQ) potentiates anti-MPM efficacy of A-1155463 in vivo. A Development of MESO-1 xenograft tumors treated with the indicated drugs. Data were presented as mean ± SEM (n = 5). #p < 0.05, comparison between combination (A-1155463 plus HCQ)- versus A- 1155463-treated group by two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, and **** p < 0.001, comparison between combination- versus HCQ-treated group by two-way ANOVA with Tukey’s multiple comparisons test. B, C Tumor size (B) and weights (C) after the treatment. Data were presented as mean ± SEM (n = 5). *p < 0.05 and ** p < 0.01 by unpaired t-test. D Mice body weights during the treatment. Data were presented as mean ± SEM (n = 5). E Immunohistochemical staining for LC3B, p62, and cleaved caspase-3 of MESO-1 xenograft tumors after the treatment. Quantification of p62 and cleaved Caspase-3 in the entire tissue sections were performed by QuPath. F Immunoblots for LC3B, p62, and cleaved caspase-7 of MESO-1 xenograft tumors after the treatment. Protein quantification is shown to the right.

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 13 of 14

apoptosis, as they accumulate sufficient amounts of the therapeutic potential of BCL-XL-specific BH3 mimetics pro-apoptotic activators36, which has engendered the in MPM, alone and in combination with the FDA- concept of cancer treatment by conquering or over- approved HCQ. whelming anti-apoptotic defenses, e.g., blockage of spe- fi Acknowledgements ci c or multiple pro-survival proteins with BH3 mimetics The results published here are in part upon data generated by the TCGA such as ABT-263 (navitoclax) and ABT-199 (veneto- Research Network: https://www.cancer.gov/tcga. Robert Kratzke (Masonic clax)12,42. In MPM, apoptosis suppression was reported to Cancer Center, University of Minnesota, USA) and Mario P. Tschan (Institute of 43 Pathology, University of Bern, Switzerland) are acknowledged for kindly be promoted by defects in core-apoptosis signaling , and providing LP-9 cells and lentiviral particles expressing mCherry-eGFP-LC3B, the pro-apoptotic BH3 mimetic ABT-737 targeting BCL- respectively. We thank Nivia Barontini (Thoracic Surgery Laboratory, Bern 44 2/BCL-XL/BCL-W and a pan-BCL-2 inhibitor (JY-1- University Hospital), Kristina Krempaska (Department for Biomedical Research, 106) were active against MPM cells45,46. However, despite University of Bern, Switzerland) for technical assistance. the promising clinical activity of pan-BH3 mimetic drugs, Author details challenges still prevail due to intrinsic or/and acquired 1Division of General Thoracic Surgery, Inselspital, Bern University Hospital, Bern, – 2 resistance and on-target platelet toxicity14,47 49, necessi- Switzerland. Department for BioMedical Research (DBMR), University of Bern, Bern, Switzerland. 3Interfaculty Bioinformatics Unit and Swiss Institute of tating the need to dissect the survival dependency on Bioinformatics, University of Bern, Bern, Switzerland. 4Institute of Pathology, individual BCL-2 proteins and the use of selective BH3 University of Bern, Bern, Switzerland mimetics in clinical development. We show here that Author contributions BCL-XL is a major survival dependency for MPM cells D.X., S.-Q.L., R.A.S., and R.-W.P. contributed to design, data acquisition, analysis, and that the BCL-XL by selective BH3 mimetic demon- and interpretation; Z.Y., H.Y., R.B., S.O., S.B., T.M.M., S.R.R.H., P.D., and G.J.K. strates therapeutic potential for subsets of MPM. Notably, contributed to analysis and interpretation of data; D.X. drafted the manuscript and R.-W.P. critically revised the manuscript. All authors read and agreed on our data are consistent with earlier observations that the final version of the manuscript. antagonizing BCL-XL by alternative strategies (e.g., anti- 44,50 sense oligonucleotides) suppresses MPM cell survival Funding and with the finding of a very recent study51 published This work was funded by grants from the Swiss Cancer League/Swiss Cancer Research Foundation (#KFS-4851-08-2019; to RWP), PhD fellowships from amid the manuscript preparation of our work. China Scholarship Council (to H.Y. and Z.Y.). We showed for the first time that BCL-XL inhibition elicits protective autophagy that limits the efficacy of Ethics statement The animal experiments were approved by the Veterinary Office of Canton BCL-XL-selective BH3 mimetics. Autophagy and apop- Bern, Switzerland. tosis constitute two important self-destructive processes 52 to maintain cellular homeostasis , and there is a complex Conflict of interest – reciprocal interplay53 55. Apoptosis activation can either The authors declare no competing interests. increase or decrease autophagy, but the underlying 56 mechanisms are controversial . Recent studies have Publisher’s note reported that inhibition of pro-survival BCL-2 proteins with Springer Nature remains neutral with regard to jurisdictional claims in BH3 mimetics could induce autophagy either by releasing published maps and institutional affiliations. Beclin-1 from the BH3-binding groove of BCL-2/BCL-XL Supplementary information The online version contains supplementary 57–59 or by BAX- and BAK1-mediated LC3B lipidation .In material available at https://doi.org/10.1038/s41419-021-03668-x. line with this notion, we revealed that targeting autop- hagy with HCQ synergistically enhances the cytotoxic Received: 13 March 2020 Revised: 22 March 2021 Accepted: 22 March 2021 effect of A-1155463, suggesting that this combination may be a novel strategy for treating MPM. Notably and in further support of our findings, a recent report pub- lished amid the revision of this study showed that BCL- References 1. Scherpereel,A.,Wallyn,F.,Albelda,S.M.&Munck,C.Noveltherapiesfor XL is overexpressed and is an important pro-survival malignant pleural mesothelioma. Lancet Oncol. 19,e161–e172 (2018). proteininMPMcells60. 2. Hmeljak, J. et al. Integrative molecular characterization of malignant pleural The lack of therapeutically exploitable mutations has mesothelioma. Cancer Disco. 8, 1548–1565 (2018). fi 3. Wu, L. et al. Progress of malignant mesothelioma research in basic science: a signi cantly hampered the development of targeted review of the 14th international conference of the international mesothelioma therapies for MPM2,6, which, however, highlights the interest group (iMig2018). Lung Cancer 127,138–145 (2019). importance to identify oncogenic dependencies rather 4. Rossini, M. et al. New perspectives on diagnosis and therapy of malignant fi 1 pleural mesothelioma. Front. Oncol. 8,91(2018). than speci c driver mutations to combat MPM . We have 5. Guo, G. et al. Whole-exome sequencing reveals frequent genetic alterations in systematically assessed the pro-survival BCL-2 proteins BAP1, NF2, CDKN2A, and CUL1 in malignant pleural mesothelioma. Cancer for their contributions to anti-apoptosis in MPM cells, Res. 75,264–269 (2015). 6. Bueno, R. et al. Comprehensive genomic analysis of malignant pleural which, to the best of our knowledge, has remained mesothelioma identifies recurrent mutations, gene fusions and splicing incompletely defined. Overall, our work demonstrates the alterations. Nat. Genet. 48,407–416 (2016).

Official journal of the Cell Death Differentiation Association Xu et al. Cell Death and Disease (2021) 12:406 Page 14 of 14

7. Vogelzang, N. J. et al. Phase III study of pemetrexed in combination with 35. Juin, P., Geneste, O., Gautier, F., Depil, S. & Campone, M. Decoding and cisplatin versus cisplatin alone in patients with malignant pleural mesothe- unlocking the BCL-2 dependency of cancer cells. Nat. Rev. Cancer 13, 455–465 lioma. J. Clin. Oncol. 21,2636–2644 (2003). (2013). 8. Li,W.,Cooper,J.,Karajannis,M.A.&Giancotti,F.G.Merlin:atumoursuppressor 36. Montero, J. & Letai, A. Why do BCL-2 inhibitors work and where should we use with functions at the cell cortex and in the nucleus. EMBO Rep. 13,204–215 them in the clinic? Cell Death Differ. 25,56–64 (2018). (2012). 37. Certo, M. et al. Mitochondria primed by death signals determine cellular 9. Yu, F.-X., Zhao, B. & Guan, K.-L. Hippo pathway in organ size control, tissue addiction to antiapoptotic BCL-2 family members. Cancer Cell 9,351–365 homeostasis, and cancer. Cell 163,811–828 (2015). (2006). 10. Benhamouche, S. et al. Nf2/Merlin controls progenitor homeostasis and 38. Klionsky, D. J. et al. Guidelines for the use and interpretation of assays for tumorigenesis in the liver. Genes Dev. 24, 1718–1730 (2010). monitoring autophagy (3rd edition). Autophagy 12,1–222 (2016). 11. Hanahan,D.&Weinberg,R.A.Hallmarksofcancer:thenextgeneration.Cell 39. Galluzzi, L. et al. Autophagy in malignant transformation and cancer pro- 144,646–674 (2011). gression. EMBO J. 34,856–880 (2015). 12. Merino, D. et al. BH3-mimetic drugs: blazing the trail for new cancer medi- 40. Liu, J. L. et al. Prognostic significance of p62/SQSTM1 subcellular locali- cines. Cancer Cell 34,879–891 (2018). zation and LC3B in oral squamous cell carcinoma. Br. J. Cancer 111, 13. Fesik,S.W.Promotingapoptosisasastrategy for cancer drug discovery. Nat. 944–954 (2014). Rev. Cancer 5, 876–885 (2005). 41. Follo,C.,Barbone,D.,Richards,W.G.,Bueno,R.&Broaddus,V.C.Autophagy 14. Leverson, J. D. et al. Exploiting selective BCL-2 family inhibitors to dissect cell initiation correlates with the autophagic flux in 3D models of mesothelioma survival dependencies and define improved strategies for cancer therapy. Sci. and with patient outcome. Autophagy 12, 1180–1194 (2016). Transl. Med. 7, 279ra240 (2015). 42.Zhang,L.,Ming,L.&Yu,J.BH3mimeticstoimprovecancertherapy; 15. Stilgenbauer, S. et al. Venetoclax in relapsed or refractory chronic lymphocytic mechanisms and examples. Drug Resist. Updat. 10,207–217 (2007). leukaemia with 17p deletion: a multicentre, open-label, phase 2 study. Lancet 43. Fennell, D. A. & Rudd, R. M. Defective core-apoptosis signalling in diffuse Oncol. 17,768–778 (2016). malignant pleural mesothelioma: opportunities for effective drug develop- 16. Broaddus,V.C.,Yang,L.,Scavo,L.M.,Ernst,J.D.&Boylan,A.M.Asbestos ment. Lancet Oncol. 5, 354–362 (2004). induces apoptosis of human and rabbit pleural mesothelial cells via reactive 44. Mohiuddin,I.,Cao,X.,Fang,B.,Nishizaki,M.&Smythe,W.R.Significant aug- oxygen species. J. Clin. Invest. 98, 2050–2059 (1996). mentation of pro-apoptotic gene therapy by pharmacologic bcl-xl down- 17. Soini, Y. et al. Apoptosis and expression of apoptosis regulating proteins bcl-2, regulation in mesothelioma. Cancer Gene Ther. 8,547–554 (2001). mcl-1, bcl-X, and bax in malignant mesothelioma. Clin. Cancer Res. 5, 45. Barbone, D. et al. The Bcl-2 repertoire of mesothelioma spheroids underlies 3508–3515 (1999). acquired apoptotic multicellular resistance. Cell Death Dis. 2, e174 (2011). 18. Xu, D. et al. Increased sensitivity to apoptosis upon endoplasmic reticulum 46. Cao,X.etal.ThenovelBH3alpha-helix mimetic JY-1-106 induces apoptosis in stress-induced activation of the unfolded protein response in chemotherapy- a subset of cancer cells (lung cancer, colon cancer and mesothelioma) by resistant malignant pleural mesothelioma. Br. J. Cancer 119,65–75 (2018). disrupting Bcl-xL and Mcl-1 protein-protein interactions with Bak. Mol. Cancer 19. Jacobson, B. A. et al. Activated 4E-BP1 represses tumourigenesis and IGF-I- 12, 42 (2013). mediated activation of the eIF4F complex in mesothelioma. Br.J.Cancer101, 47. Faber, A. C. et al. mTOR inhibition specifically sensitizes colorectal cancers with 424–431 (2009). KRAS or BRAF mutations to BCL-2/BCL-XL inhibition by suppressing MCL-1. 20. Xu, D. et al. Endoplasmic reticulum stress signaling as a therapeutic target in Cancer Disco. 4,42–52 (2014). malignant pleural mesothelioma. Cancers 11, 1502 (2019). 48. Opferman, J. T. Attacking cancer’s Achilles heel: antagonism of anti-apoptotic 21. Chou, T. C. Drug combination studies and their synergy quantification using BCL-2 family members. FEBS J. 283,2661–2675 (2016). the Chou-Talalay method. Cancer Res. 70,440–446 (2010). 49. Budhraja, A. et al. Modulation of navitoclax sensitivity by dihydroartemisinin- 22. Xu, D. et al. CRISPR screening identifies WEE1 as a combination target for mediated MCL-1 repression in BCR-ABL(+) B-lineage acute lymphoblastic standard chemotherapy in malignant pleural mesothelioma. Mol. Cancer Ther. leukemia. Clin. Cancer Res. 23, 7558–7568 (2017). 19,661–672 (2020). 50. Hopkins-Donaldson, S. et al. Induction of apoptosis and chemosensitization of 23. Schneider,C.A.,Rasband,W.S.&Eliceiri,K.W.NIHimagetoImageJ:25years mesothelioma cells by Bcl-2 and Bcl-xL antisense treatment. Int. J. Cancer 106, of image analysis. Nat. Methods 9,671–675 (2012). 160–166 (2003). 24. Schlafli,A.M.,Berezowska,S.,Adams,O.,Langer,R.&Tschan,M.P.ReliableLC3 51. Jackson,M.R.etal.Mesothelioma cells depend on the antiapoptotic protein and p62 autophagy marker detection in formalin fixed paraffin embedded Bcl-xL for survival and are sensitized to ionizing radiation by BH3-mimetics. Int. human tissue by immunohistochemistry. Eur. J. Histochem. 59,2481(2015). J. Radiat. Oncol. Biol. Phys. 106,867–877 (2020). 25. Bankhead, P. et al. QuPath: open source software for digital pathology image 52. Galluzzi, L. et al. Molecular mechanisms of cell death: recommendations of the analysis. Sci. Rep. 7, 16878 (2017). Nomenclature Committee on Cell Death 2018. Cell Death Differ. 25,486–541 26. Parejo, S. et al. Assessing autophagy during retinoid treatment of (2018). cells. Methods Mol. Biol. 2019,237–256 (2019). 53. Maiuri, M. C., Zalckvar, E., Kimchi, A. & Kroemer, G. Self-eating and self-killing: 27. Kinsey, C. G. et al. Protective autophagy elicited by RAF->MEK->ERK inhibition crosstalk between autophagy and apoptosis. Nat. Rev. Mol. Cell Biol. 8, 741–752 suggests a treatment strategy for RAS-driven cancers. Nat. Med. 25,620–627 (2007). (2019). 54. Marino, G., Niso-Santano, M., Baehrecke, E. H. & Kroemer, G. Self-consumption: 28. Yang, H. et al. /AXL/eIF4E-regulated unfolded protein response as an the interplay of autophagy and apoptosis. Nat. Rev. Mol. Cell Biol. 15,81–94 acquired vulnerability in drug-resistant KRAS-mutant lung cancer. Oncogenesis (2014). 8, 45 (2019). 55. Fitzwalter, B. E. et al. Autophagy inhibition mediates apoptosis sensitization in 29. Li, J. et al. TCPA: a resource for cancer functional proteomics data. Nat. cancer therapy by relieving FOXO3a turnover. Dev. Cell 44,555–565 (2018). e3. Methods 10,1046–1047 (2013). 56. Gump, J. M. & Thorburn, A. Autophagy and apoptosis: what is the connection? 30. Cerami, E. et al. The cBio cancer genomics portal: an open platform for Trends Cell Biol. 21,387–392 (2011). exploring multidimensional cancer genomics data. Cancer Disco. 2,401–404 57. Pedro, J. M. et al. BAX and BAK1 are dispensable for ABT-737-induced dis- (2012). sociation of the BCL2-BECN1 complex and autophagy. Autophagy 11, 31. Tsherniak,A.etal.Defining a cancer dependency map. Cell 170,564–576 452–459 (2015). (2017). e516. 58. Reljic, B. et al. BAX-BAK1-independent LC3B lipidation by BH3 mimetics is 32. Barrett, T. et al. NCBI GEO: archive for functional genomics data sets—update. unrelated to BH3 mimetic activity and hasonlyminimaleffectsonautophagic Nucleic Acids Res. 41,D991–D995 (2012). flux. Autophagy 12,1083–1093 (2016). 33. Yang, W. et al. Genomics of drug sensitivity in cancer (GDSC): a resource for 59. Lin,Q.H.etal.ABT-263inducesG1/G0-phase arrest, apoptosis and autophagy therapeutic biomarker discovery in cancer cells. Nucleic Acids Res. 41, in human esophageal cancer cells in vitro. Acta Pharm. Sin. 38,1632–1641 D955–D961 (2013). (2017). 34. Tate, J. G. et al. COSMIC: the catalogue of somatic mutations in cancer. Nucleic 60. Arulananda, S. et al. BCL-XL is an actionable target for treatment of malignant Acids Res. 47,D941–D947 (2019). pleural mesothelioma. Cell Death Disco. 6, 114 (2020).

Official journal of the Cell Death Differentiation Association